React Patterns
Security checks across malware telemetry and agentic risk
Overview
The available evidence shows a clean, documentation-style skill review with no artifact-backed security concern.
This appears safe to install based on the available evidence. As with any skill, review the actions your agent proposes before allowing code changes, network calls, credential use, or production-impacting operations.
SkillSpector
By NVIDIA
Vulnerability Patterns
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
VirusTotal
65/65 vendors flagged this skill as clean.